Intelligent Infrastructure and Construction

Conventional infrastructure asset management relies heavily on continuous, linear degradation models, which often fail to capture sudden macroscopic failures. This study computationally explores these limitations by investigating non-linear degradation processes under specified boundary conditions. We propose a stochastic differential equation (SDE) framework incorporating jump processes and specular reflection boundaries to model the macroscopic irreversibility of structural decay. To illustrate this framework’s behavior, large-scale Monte Carlo simulations involving up to 1,000,000 cohorts were conducted. The computational results numerically demonstrate that, entirely by model construction, continuous drift and diffusion are explicitly prevented from breaching the imposed reflection boundaries. Consequently, absorption at the failure state is driven exclusively by the jump term. Rather than claiming this as an absolute physical necessity or a mathematical proof of real-world degradation, this study aims to computationally illustrate the macroscopic consequences of imposing such strict boundary rules. Furthermore, the simulations show that under these specific reflection constraints, environmental noise cannot mathematically extend the system lifespan—a restricted dynamic we call “negative resilience.” Acknowledging the absence of empirical validation and alternative boundary comparisons, these exploratory findings serve as a theoretical baseline to understand boundary-driven jump-diffusion behavior, highlighting the inherent limitations of relying purely on continuous predictive models.

This study presents an Internet of Things (IoT)-enabled Structural Health Monitoring (SHM) system for the real-time monitoring of excavation-induced movements in shoring walls using a Micro-Electro-Mechanical Systems (MEMS)-based wireless inclinometer. Conventional inclination measurements were impractical due to restricted site accessibility, steep excavation slopes, rock-fall hazards, confined working spaces, and safety risks. To overcome these limitations, a custom-designed, ruggedised monitoring system integrated with a cloud-based Graphical User Interface (GUI) was developed to enable continuous remote acquisition, visualisation, and analysis of inclination data. A 65-day field investigation was conducted on two shoring walls at a deep excavation site to evaluate the influence of construction activities and environmental conditions on wall stability. The monitoring system successfully captured inclination variations induced by drilling operations, physical shocks, and rainfall, demonstrating its capability to provide continuous assessment of structural behaviour under dynamic site conditions. The proposed IoT framework enabled early detection of abnormal wall movements and facilitated timely engineering intervention, thereby reducing operational risks and improving site safety. The developed system provides a cost-effective and reliable solution for real-time monitoring of temporary excavation support systems and offers significant potential for enhancing decision-making, risk mitigation, and predictive maintenance in geotechnical and structural engineering applications.

The adoption of robotics in construction has progressed more slowly than in many other sectors. Building information models (BIMs) contain information that can support construction robotics, while robots require task-specific project data, such as geometry, location, material, sequence, and operating constraints, that may not be readily available in a directly usable form. This study proposes a conceptually integrated system architecture for BIM-to-robot information exchange comprising Task Planning, Task Decomposition, and anticipated Robot Task Execution linked by two information exchanges. A 20-concrete-masonry-unit (CMU) wall case study demonstrates selected information-preparation and documentation portions of the architecture. A low-LOD Revit wall was decomposed into individual CMUs using Dynamo, and component location (X, Y, Z, and orientation), material type, and identification (I.D.) information were extracted and formatted as robot-oriented task information. Because the available robotic platform lacked manipulation capability, a human worker interpreted the exported task information and constructed the wall, while a Husky A200 robot collected site images for dense point-cloud reconstruction. The reconstruction was superimposed on the source Revit model to qualitatively assess consistency between the constructed wall and the intended BIM-derived layout. The case study therefore demonstrates BIM-derived information generation and extraction, worker interpretation and use of the task information, and robot-assisted documentation, but it does not verify autonomous robotic construction. The primary contribution is the architectural integration of prior information-exchange and parametric-modeling methods across the end-to-end BIM-to-robot workflow, including explicit information requirements and exchange points; the term end-to-end refers to architectural scope rather than end-to-end experimental validation.